[DigitalToday reporter Chi-gyu Hwang (황치규)] Nvidia said on Aug. 25 it unveiled the Jetson Orin Nano 2, an entry-level edge AI computer for robotics.

The company said the product is designed to run AI and frontier-class models inside robots and drones.

Compute performance is twice that of the previous generation. It has 8 gigabytes of memory and an 8-core central processing unit. Power consumption is down 40 percent in 15-watt power mode.

Nvidia said AI models have recently become smaller and more efficient, making it possible to deliver language understanding, computer vision and audio processing on smaller hardware. It said it has improved the hardware so smaller, faster and denser models can run efficiently at the edge.

Deepu Talla (디푸 탈라), Nvidia vice president for robotics and edge AI, said small and mid-sized frontier models have reached the accuracy of last year’s largest frontier models, enabling real-time intelligence on edge devices.

The Jetson Orin Nano 2 is designed to combine high processing performance and power efficiency in a small form factor.

Drones and lightweight robots can run real-time vision systems directly on devices, reducing the computing burden previously offloaded to the cloud. That allows them to respond immediately without waiting for results to be returned over wireless communications.

Nvidia said Cognex, Doosan Bobcat, Matic Robotics and Wing Aviation are working to apply the Jetson Orin Nano 2 to smart robotics. Wing Aviation is reviewing plans to use the product for its autonomous drone delivery fleet. It plans to use the Jetson Orin Nano 2 and its software stack to improve real-time AI perception performance needed for autonomous navigation, avoidance and inference.

Dinuka Abeywardena (디누카 아베이워데나), head of perception at Wing Aviation, said it would help develop more responsive and energy-efficient drones. Nvidia said it plans to supply the Jetson Orin Nano 2 to the public in module and developer kit forms starting in the first half of 2027.